Figure legends
Figure 1. Operational settings chosen, with the measured feast length
during the start-up period for obtaining granules.
Figure 2. Visual impression of the sludge bed during the different
operational periods in this study. From left to right, Start-up phase,
40-50 d SRT, 10-20 d SRT, and 1-2 d SRT.
Figure 3. Upper picture is an overview of the reactor performance in
terms of VFA, VSS production and the compactness of the culture
expressed as SVI60. The bottom picture is the evolution of the product
spectrum obtained during each enrichment. Day 0-136 is the 40-50 d SRT,
day 137-192 is the 10-20 d SRT and finally day 193-212 is the 1-2 d SRT
system.
Figure 4. All system characterisations performed in this study. From top
to bottom the 40-50 d SRT, 10-20 d SRT and finally the 1-2 d SRT is
depicted. Markers reflect measurements and lines are modelled. Soluble
COD concentrations (sCOD) were calculated as the sum of the individual
substrate and product concentrations. All cycle measurements were
performed on the last day of the enrichment thus on days: 136, 192 and
212 for respectively 40-50 d, 10-20 d and 1-2 d SRT.
Appendix A
The mathematical model made in this study was aimed to derive
characteristic parameters from the experimental data. The model was
based upon 3 reactions: (i) the direct conversion of glucose to organic
acids, storage polymers, and biomass; (ii) the conversion of storage
polymers to organic acids and biomass; (iii) when lactate was one of the
organic acids produced a third reaction was used describing the
conversion of lactate to acetate, propionate and biomass. Glucose and
lacatate conversions occurred in parallel and resulting in (partial)
similar products thus these could not be calibrated separately. For this
reason the reaction stoichiometry for lactate was fixed and similar to
that used by (J. Tamis et al., 2015) and shown in table A.1. The
reaction stoichiometry for reaction (i) and (ii) are given in table A.1
Table A.1. The reaction stoichiometries used to calibrate the model to
the measured data